Hawaiian Lowland Wet Forests: Impacts of Invasive Plants on Light Availability

Author:  Christina P. Wong

Institution:  Occidental College
Date:  May 2007


The Hawaiian Islands contain a variety of climatic regions, elevations and substrates, which are home to a large number of endemic flora and fauna. Unfortunately, these rare ecosystems and endemic species are facing a bleak future of endangerment and extinction, such as the Hawaiian lowland wet forest. Non-native and invasive plant and animal populations are currently out-competing and over-crowding native lowland species. This study evaluated: (i) the effect of alien plants on lowland canopy structure and light availability and (ii) the resource efficiency of several dominant native and alien trees. Removal plots determined that alien species are altering under story light availability by creating a dense (1-10m) sub-canopy that is limiting forest light heterogeneity, and decreasing light transmittance by 51%. Light response curves conducted on Psychotria hawaiiensis (an endemic species) and two alien species (Macaranga mappa and Melastoma candidum) show that Psychotria has lower photosynthesis rates, but similar light compensation and saturation points, thus having shade-tolerant capabilities similar to M. mappa and M. candidum. However, the two alien species demonstrated higher specific leaf areas and lower leaf construction costs. This study indicates that P. hawaiiensis may be the most prominent endemic sub-canopy species because of its competitive photosynthetic characteristics. Yet, P. hawaiiensis constitutes only a small percentage of the total sub-canopy likely because M. mappa and M. candidum have faster growth and regeneration periods. These results suggest that alien plant species are drastically affecting forest under story light availability because of their leaf allocation patterns and high growth rates.

Symbols:(native), Non-Nat. (non-native), PPFD (percentage of photon flux density), PFD (photon flux density), SLA (specific leaf area), CC (leaf cost of construction), and A (CO2 assimilation).


Being the most secluded archipelago on Earth, Hawaii's ecosystems are extremely susceptible to biological invasions (Vitousek, 1995). Its native plant populations evolved in an environment absent of grazing mammals and fierce plant competitors making them vulnerable to extinction by newly introduced alien species (Sakai et al., 2002). Define invasive species, not very clear.' Invasive species can dramatically alter ecosystem-level processes by affecting resource availability, trophic structure, and disturbance intensity (Vitousek 1995). Hawaii, currently houses several thousand known alien plant species of which several present serious threats to native ecosystems (Vitousek, 1995).

One of the most threatened Hawaiian ecosystems is lowland wet forests, yet little is known of their Ecology and historic species composition. These forests reside at

< 800m and whose annual rainfall is > 2500 mm at 1000m and > 3000mm at sea level (Price et al. In Review). Lowland wet forests were once abundant throughout the Hawaiian Islands but now reside primarily on the Eastern side of the Big Island of Hawaii (Ralph, 1982; Ziegler, 2002). Lowland areas are highly developed and disturbed making them susceptible to rapid invasions. They were first altered for agricultural purposes by the Polynesians and continue to be cleared for urban expansion and tourism (Ziegler, 2002). Today, the forest structure no longer resembles its native past with 50% of its plant taxa at risk of endangerment and 8.33% declared extinct (Sakai et al., 2002). Preservation of Hawaii's lowland wet forest is dependent upon our understanding of its current ecological condition.

Native Hawaiian forests are relatively broken or open compared to continental tropical forests; thus many dominant native species need high light environments for germination and survival (Burton and Mueller-Dombois, 1984; Drake and Mueller-Dombois, 1993). Non-native invaders seem to capitalize on these weaknesses by possessing adaptations that allow successful germination and propagation in high and low light communities (Baruch et al., 2000; DeWalt et al., 2004). Several studies suggest that the key mechanisms for successful invasion are increased resource efficiency, differences in allocation and investment (e.g., decreased leaf cost of construction (CC) and increased specific leaf area (SLA)), and higher and more effective levels of reproduction (Pattison et al., 1998; Baruch and Goldstein, 1999; Baruch et al., 2000). Species variation towards light acquisition is a major factor in determining the effect alien species may have on native habitats.

This study examined the effects of alien species on canopy structure, light dynamics and resource efficiency of three sub-canopy and understory species: Psychotria hawaiiensis (endemic), Macaranga mappa and Melastoma candidum, (both invasive), in a lowland wet forest in Hilo, HI. This study asked three questions. First, how are invasive plant species altering canopy structure, light heterogeneity, and light availability in the forest understory? Second, how does plant growth and carbon assimilation rates differ between native and non-native species based on allocation patterns? Thirdly, how can P. hawaiiensis survive in a highly invaded under story? It is predicted that invasive plant species were decreasing Hawaiian lowland wet forest light heterogeneity by creating a dense sub-canopy of alien vegetation. Furthermore, it hypothesized that M. mappa and M. candidum would demonstrate shade-tolerant leaf characteristics, higher photosynthesis rates, and lower leaf construction costs than P. hawaiiensis.

Materials and Methods

Site History and Current Forest Structure

The site was located in the Keaukaha Military Reservation (KMR) in Hilo, Hawaii. Approximately 100 acre fence was erected in 2002 to exclude ungulates. The lowland forest structure was comprised of a canopy, sub-canopy and under story. The canopy was > 20m tall and primarily housed Metrosideros polymorpha (Native), Diospyros sandwicensis (Native), and Cecropia obtusifolia (Non-Native). The sub-canopy was composed of mainly Melastoma candidum (Non-Native), Psidium cattleianum (Non-Native), Pandanus tectorius (Native), Psychotria hawaiiensis (Native), Macaranga mappa (Non-Native), Alyxia oliviformis (Non-Native) and Paederia foetida (Non-Native). The forest understory predominately consisted of M. candidum, P. hawaiiensis, and M. mappa saplings, Clidemia hirta (Non-Native) and ferns (Wagner et al., 1999).

Eight paired-plot comparisons had been previously created at the site in early 2004: four removal plots (native plant species only) and four adjacent control plots (non-native + native plant species) along the Puna Trail in KMR. The plots were 10 x 10 m with a 5 m buffer to prevent edge effects with each paired plot at least 20 m apart.

Vertical Leaf Profile

Both canopy stratification and vertical composition by creating a vertical leaf profile were measured (Montgomery and Chazdon 2001). The presence/absence of vegetation was recorded using a telescoping pole at different heights: 0-1, 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-10 and > 10m. For each plot three intervals were marked at 3m, 6m, and 9m along the north-south direction of each plot. At each interval ten measurements (n = 30) were taken every meter along the west-east direction.

Light Availability

Photosynthetic photon flux density (PPFD) measurements were made under overcast conditions to limit the effects of solar angle and sun flecks. PPFD measurements within plots represented below-canopy readings while above-canopy measurements were taken simultaneously in a nearby open field by a quantum sensor mounted to a tripod. The sensor was connected to a LAI-2000 datalogger (LICOR Inc., Lincoln NE) that recorded PPFDs every 30 seconds. Light transmittance was a ratio of below-canopy: above-canopy PPFD measurements at two different heights: i) 1 meter (understory) and ii) 6-8m (above sub-canopy) using a quantum sensor and a LAI-2000 data logger (LICOR Inc.) (Montgomery and Chazdon, 2001). The 6-8m recordings were made by mounting the quantum sensor to a small, self-leveling platform atop a telescoping pole. At each height level ten PPFD measurements were made at three intervals: 3m, 6m, and 9m along the north-south direction of each plot. Each measurement was taken every meter along the west-east direction (identical to vertical leaf profile methodology, N = 30). Leaf area index (LAI) calculated the percentage diffuse transmittance of both 1m and 6-8m heights. The LICOR C2000 software (LICOR Inc.) computed LAI by matching the below-canopy measurements with the closest-in-time above-canopy readings.

Photosynthesis Measurements

Leaf carbon assimilation (A) was measured in the field, from 0800 to 01230 hours, with a LICOR 6400 portable gas exchange system (LICOR Inc.). Measurements were made on randomly chosen young fully developed leaves of: P. hawaiiensis (N = 6), M. mappa (N = 5) and M. candidum (N = 6). After measurements were taken leaves were collected for further analysis. Light response curves were created at a range of PPFDs from 1500 to 0 μmol m-2 s-1, 60-80% relative humidity, 25 °C block temperature, and 400 μmol m-1 CO2. Light saturated (maximum) photosynthetic rate (Amax), light saturation point (light level at maximum rate of photosynthesis), and light compensation (light level at which CO2 assimilation is zero) point were estimated for each species using Photosyn Assistant (Dundee Scientific, Dundee UK), which determines these parameters by fitting the light response data to a model function expressed as a quadratic equation (Prioul and Chartier, 1977). The variables are estimated using least squares fitting regression and the Nelder-Mead minimization routine (Nelder and Mead, 1965).

Leaf Morphology: Nutrient levels and Energy Concentration

Leaf area and dry weight were determined for all collected leaves. Leaf area was calculated using the LI-3100 (LICOR Inc.) area meter. Specific leaf area (SLA), an indicator of leaf thickness, was calculated for each species (Chiariello et al. 1991). Leaf carbon (C) and nitrogen (N) concentrations were determined by elemental analysis, ECS 4010 (Costech Analytical Technologies Inc., Valencia CA). Leaf construction costs (CC) were calculated from N (considering ammonium as the source of N), energy, and ash concentrations using Parr 1425 Semimicro bomb calorimeter (Parr Instrument Company, Moline IL), following the equations in Williams et al. (1987). Leaf CC was calculated on both a mass and an area basis (Baruch and Goldstein 1999).

Statistical Analysis

Statistics were conducted with the JMP IN v 3.17 (SAS Institute Inc., Cary NC) statistical program. All data were non-normal and thus nonparametric Wilcoxon / Kruskal-Wallis tests were used to evaluate percent light transmittance by height and treatment. Multiple comparisons were analyzed using a Tukey test, which compared light saturation and compensation points, specific leaf areas and leaf construction costs amongst species.


Vertical Leaf Profile

Foliage profiles showed distinct differences in vertical vegetation between the control and removal plots (Fig. 1A; Fig. 1B). The mean percentage of vertical foliage in the control plots varied from 27.45%- 69.95% (Fig. 1A) while the removal plots varied from 3.7%- 41.6% (Fig. 1B). Control plots had significantly more vegetation cover at all intervals ranging from 1-10m (P-value 10m did not significantly differ amongst treatments.

Figure 1. Foliage height profiles in four control plots (A) and in four removal plots (B) at Keaukaha Military Reservation's Hawaiian lowland wet forest in Hilo, HI. The asterisks indicate significant difference between control and removal plots (nonparametric Wilcoxon test: P-value < .0001). Data are means and standard error bars.

Figure 1. Foliage height profiles in four control plots (A) and in four removal plots (B) at Keaukaha Military Reservation's Hawaiian lowland wet forest in Hilo, HI. The asterisks indicate significant difference between control and removal plots (nonparametric Wilcoxon test: P-value < .0001). Data are means and standard error bars.

Light Availability and Forest structure

Mean percentage of diffuse light transmittance (%T) differed significantly between treatment plots by height (Fig. 2). In the forest understory the removal plots received 14.8 times more PPFD than the control plots.

Photosynthetic Characteristics of Species: Light Response Curves

There was a significant difference in the maximum CO2 assimilation (Amax) rate of M. mappa, and P. hawaiiensis, but M. candidum did not differ from the other species (Table 1). In contrast, there were no significant differences in their mean light compensation points and mean light saturation points of any species (Table 1). All light compensation points were between 3-5 μmol m-2 s-1 the range for shade-tolerant species (1-5 μmol m-2 s-1) (Taiz and Zieger 1991). M. candidum and M. mappa samples showed great variability in their light compensation points (3.43-6.43 μmol m-2 s-1; 3.38-6.53 μmol m-2 s-1) and light saturation points (64.9-97.2 μmol m-2 s-1; 66.4-116 μmol m-2 s-1 ), whereas, P. hawaiiensis light compensation points (3.22-5.43 μmol m-2 s-1) were relatively consistent.

Figure 2. By treatment mean percent light transmittance (± standard error bars) at 1m and 6-8m heights. For both heights light transmittance between control and removal plots were significantly different (nonparametric Wilcoxon test: P-value < 0.05).

Figure 2. By treatment mean percent light transmittance (± standard error bars) at 1m and 6-8m heights. For both heights light transmittance between control and removal plots were significantly different (nonparametric Wilcoxon test: P-value < 0.05).

Leaf Morphology/ Leaf Cost of Construction

The mean SLA of P. hawaiiensis was significantly lower than M. candidum and M. mappa. In comparison to both alien species P. hawaiiensis had a higher CCmass and CCarea. The CCarea of P. hawaiiensis was roughly double that of M. candidum and approximately 1.5 times greater than M. mappa. All species showed a common trend of increasing SLA with decreasing CCarea.

Table 1. Average specific leaf area (SLA), % leaf nitrogen concentration on mass basis (Nmass), leaf construction costs on mass (CCmass) and area basis (CCarea), maximum net CO2 assimilation (Amax), light compensation points and light saturation points. P-values for comparisons amongst all three species; nonparametric Wilcoxon test. * = Statistically significant difference; different letters indicate significant differences amongst species.

Table 1. Average specific leaf area (SLA), % leaf nitrogen concentration on mass basis (Nmass), leaf construction costs on mass (CCmass) and area basis (CCarea), maximum net CO2 assimilation (Amax), light compensation points and light saturation points. P-values for comparisons amongst all three species; nonparametric Wilcoxon test. * = Statistically significant difference; different letters indicate significant differences amongst species.


The primary objectives of this study were: (i) assess the effects of alien plants on lowland canopy structure and light availability and (ii) gain a general understanding of the resource efficiency of native plants relative to their alien counterparts. The results suggest that the two dominant alien plants at this site are likely more efficient at utilization of resources, which appear to promote their rapid growth and increased regeneration. Overall, these data elucidate the significant and prominent role of alien species on the alteration of forest light dynamics.

Light Availability/Light Quality

Canopy structure is defined as the amount, quality and organization of aboveground plant material. Canopy structure affects environmental factors such as air temperature, leaf temperature, atmospheric moisture, soil evaporation, soil temperature, precipitation interception, duration of leaf wetness, etc. (Norman and Campbell 1991). The forest's vertical vegetation composition was inversely related to light availability in both the control and removal plots. The control plots contained roughly 25% more vegetation whereas light transmittance in the removal plots was approximately 28% greater than the controls.

The foliage profiles showed the presence of a dense alien subcanopy in the control plots at 1-10m. This conclusion was further supported with a low mean PPFD value (16.76%) at the forest understory, and the substantially higher mean PPFD value (41.44%) in the removal plots. The light reaching the understory in the removal plots was heterogeneous whereas the light in the control plots was more homogenous due to the thick blanket of alien vegetation. In the control plots the mean light transmittance variance was 2.9% while in the removals it was 37.9%, further supporting the notion that alien plants are limiting light heterogeneity of native forests. Commonly, light heterogeneity in the understory of a tropical rain forest is measured by canopy gaps created by treefalls, which are important for growth and reproduction (Denslow et al. 1990). Alien plants are potentially outcompeting native species for canopy gaps thus limiting light availability to native saplings and seedlings. Traditionally the seedlings of Metrosideros polymorpha, the most dominant native canopy species, can potentially survive up to two years in deep shade, yet depend upon canopy gaps or diebacks to maintain themselves in mature forests (Burton and Mueller-Dombois 1984). A healthy forest needs direct light (sunflecks) and an increase in red light in the forest understory as a way to promote plant regeneration (Turton 1992). Our removal plots suggest a dramatic change in both quantity and spectral quality of light reaching the forest floor.

Photosynthesis, Shade-tolerance, SLA and Leaf Construction Costs

There were no substantial differences in the photosynthetic rates of P. hawaiiensis, M. candidum, and M. mappa. P. hawaiiensis demonstrated equivalent shade-tolerant traits as M. candidum and M. mappa. All three species fell in the range of shade-tolerance. P. hawaiiensis may be one of the most abundant native subcanopy species because of its competitive shade-tolerant characteristics. Possibly, P. hawaiiensis was less abundant in the past, yet the reduction in light availability is now allowing this species to increase. Likely species diversity of native lowland flora was historically higher because light conditions were more dynamic permitting native shade-intolerant species to survive and persist. M. candidum and M. mappa had higher standard deviations for their light compensation points than P. hawaiiensis, suggesting that M. candidum and M. mappa are more phenotypically plastic, which allow them to quickly adapt to a variety of light conditions.

Specific leaf area is a plant trait that appears to be important in the regulation and control of plant functions because leaves that have a high SLA can produce large or more assimilatory surfaces for a given amount of carbon fixed (Chiariello et al. 1991). Also SLA is positively related to plant growth and efficiency because it relates to leaf thickness and longevity (Baruch and Goldstein 1999). P. hawaiiensis produces thicker leaves, and a smaller SLA, than M. candidum and M. mappa, which relates to its higher leaf construction costs. Leaf construction cost is a measure of the energy invested by plants to synthesize carbon skeletons and nitrogenous compounds, which can also be indirectly related to resource efficiency. M. candidum can produce twice the number of P. hawaiiensis leaves with the same amount of energy. M. candidum and M. mappa were more efficient at resource utilization than P. hawaiiensis, which further supports the premise that these alien species are outcompeting native plants for canopy gaps. Resource efficiency of alien species may be one of the main mechanisms for successful invasion of lowland areas.

Hawaiian lowland wet forests are extremely important because they are home to several native plants and birds that are significant cultural, social and scientific resources (Wagner et al. 1999). For example, the low elevation amakihi, a native Hawaiian honeycreeper, is showing increased abundance in lowland wet forests, suggesting potential signs of immunity to avian malaria, which could be an evolutionary breakthrough for the survival of native birds (Woodworth et al. 2004, Spiegel et al. 2006). Yet their survival is dependent upon the presence of intact native forests. Hawaii's native lowland biota are trying to survive and sustain themselves, and further research will help to understand and perpetuate the belief that Hawaii's lowland wet forests are not gone.


I thank Sharon Ziegler-Chong, H. Erika Perry, and Carmen Perez-Frayne, and Donald Price for affording me with this amazing opportunity. Funding was provided by a National Science Foundation Research Experience for Undergraduates Site Program to the University of Hawaii at Hilo (DEB # 0139379, Donald Price, PI and Sharon Ziegler-Chong, co-PI). I thank John Nygaard and Colleen Cole for their assistance in the field. Equipment and lab access were provided by the University of Hawaii-Hilo and the USDA Forest Service, Institute of Pacific Islands Forestry. Nutrient analyses were provided by Randi Schneider at the EPSCoR Analytical Lab Facility, University of Hawaii-Hilo. Access to field sites was provided by the Hawaii Army National Guard Environmental Office (Col. O. Peterson), and Cynthia Thurkins and Major Lindsey facilitated this research at Keaukaha Military Reservation


Baruch, Z. et al. (2000) Responses to light and water availability of four invasive melastomataceae in the Hawaiian Islands. International Journal Plant Science 161, 107-118.

Baruch, Z., and G. Goldstein (1999) Leaf construction cost, nutrient concentration and net CO2 assimilation of native and invasive species in Hawaii. Oecologia 121, 183-192.

Burton, P.J. and D. Mueller-Dombois (1984) Response of Metrosideros polymorpha seedlings to experimental canopy opening. Ecology 65, 779-791.

Chiariello, N.R. et al. (1991) Plant Physiological Ecology: Growth, carbon allocation and cost of plant tissues. Chapman and Hall, USA, 327- 365.

Denslow, J.S., and S.G. Guzman (2000) Variation in stand structure, light and seedling abundance across a tropical moist forest chronosequence, Panama. Journal of Vegetation Science 11, 201-212.

Denslow, J.S. et al. (1990) Growth responses of tropical shrubs to treefall gap environments. Ecology 71(1): 165-179.

DeWalt, S.J. et al. (2004) Biomass allocation, growth, and photosynthesis of genotypes from native and introduced ranges of tropical shrub Clidemia hirta. Oecologia 138, 521-531.

Drake, D.R. and D. Mueller-Dombois (1993) Population development of rainforest trees on a chronosequence of Hawaiian lava flows. Ecology 74, 1012-1019.

Montgomery, R.A., and R.L. Chazdon (2001) Forest structure, canopy, architecture and light transmittance in tropical wet forests. Ecology 82, 2707-2718.

Nelder, J. A., and R. Mead (1965) A simple method for function minimization. Computer Journal 7, 308-313.

Norman J.M., and G.S. Campbell (1991) Plant Physiological Ecology: Canopy Structure. Chapman and Hall, USA, 301-325.

Pattison, R.R. et al. (1998) Growth, biomass allocation and photosynthesis of invasive and native Hawaiian rainforest species. Oecologia 117, 449-459.

Pearcy R.W., and D.A. Sims (1994) Photosynthetic acclimation to changing light environments: scaling from the leaf to the whole plant. San Diego: Academic Press, 145-174.

Price, J.P. et al. In review. Mapping plant species ranges in the

Hawaiian Islands: Developing a methodology and associated GIS layers. HCSU

Technical Reports, Hawaii Cooperative Studies Unit, University of Hawaii at

Hilo, Pacific Aquaculture and Coastal Resources Center, Hilo, HI.

Prioul, J. L., and P. Chartier (1977) Partitioning of transfer and carboxylation components of intracellular resistance to photosynthetic CO2 fixation: A critical analysis of the methods used. Annals of Botany 41, 789-800.

Ralph, C.J. (1982) Hawaii showcase of evolution: birds of the forest. Natural History 91, 1-44.

Sakai, A.K. et al. (2002) Patterns of endangerment in the Hawaiian flora. Systematic Biology 52, 276-302.

Taiz L., and E. Zeiger (1991) Plant Physiology. California: The Benjamin/Cummings Publishing Company, Inc.

Spiegel, C. S. et al. (2006) Distribution and abundance of forest birds in low-altitude habitat on Hawaii Island: evidence for range expansion of native species. Bird Conservation International 16, 175-185.

Turton, S.M. (1992) Understory light environments in a north-east Australian rainforest before and after a tropical cyclone. Journal of Tropical Ecology 8, 241-252.

Vitousek, P. M. (1995) The Hawaiian Islands as a model system for ecosystem studies. Pacific Science 49, 2-16.

Wagner W.L. et al. (1999) Manual of Flowering Plants of Hawaii (volume #1 and 2). Honolulu: University of Hawaii Press/ Bishop Museum Press.

Woodworth, B.L. et al. (2005) Host population persistence in the face of introduced vector-borne diseases: Hawaii amakihi and avian malaria. PNAS 102, 1531-1536.

Williams K. et al. (1987) Estimation of tissue construction cost from heat of combustion and organic nitrogen content. Plant, Cell and Environment 10, 725-734.

Ziegler, A.C. (2002) Hawaiian Natural History, Ecology and Evolution. University of Hawai'i Press.